Tag Archives: AI in publishing

AI and the Future of Children’s Publishing

Our charity’s publishing imprint, Little People Books, explored how ai could support our work in children’s publishing. We used AI to generate stories with scant prompts and first drafts. We checked readability for different age groups, rendered illustrations, experimented with soundscaping and video editing for exhibitions and our YouTube channel.

What we’ve learned is that AI is a powerful tool for speed and volume and with enough character-style input from both authors and illustrators alike, it can evolve plots, navigate complex plot dynamics and multi-edit for different audiences
Summary: Once story characters are well defined, AI can alert authors when plot overlaps work and where they may need reviewing and follow character trait settings to produce engaging turns of phrase. As educational story creators, Little People Books found it useful that AI could fact check during realtime writing.

This report shares our first ai journey. What worked, what didn’t. It includes our thoughts on where the integration of ai into our charity's publishing might add, subtract and align us better with our work. We’d love to hear your thoughts and ideas on ai as we all move forward together with its developments.

AI in Children’s Publishing  – a WOW Report

Storytelling for young children, environment, and wellbeing

As editor at Little People Books,  I am excited to present the findings of our comprehensive four month study into the Utilization of AI in Publishing, with a particular focus on illustrated storybooks for young children.

Commissioned by our parent charity, World of Water, this exploration covered critical subjects including environmental awareness, personal wellbeing, literacy and sustainability.

(Above) Result of AI prompt for a back cover spread of 'The Gathering'  (Author: Paul Stanley)

(Below) AI version of data hub for IWAC [International Water Awareness Centre]

Using AI for Text, Images, Video and Sound

The project commenced with rigorous text-to-text research with ChatGPT. This initial phase provided a solid foundation, enabling us to craft engaging narratives to resonate with a variety of young audiences – each age group being analysed using ai rather than human focus groups.

For example, we were able to generate a new aspect to an existing story using this minimal prompt (reverse engineered from “Me and My Family”):

“Write a story about two fun fictitious characters living for the chance of appearing in their own book whilst having to work on producing graphic novels full of other less interesting OC’s”.

We were then able to paste the story into a series of other ai programme to get Readability Grade Levels and Readability Scores. For example, the story generated by ai for the above prompt scored the following:

Readability Grade Levels:
Flesch-Kincaid Grade Level 14.3
Gunning Fog Index 16.3
Coleman-Liau Index 16.3
SMOG Index 15.8
Automated Readability Index 14.4
FORCAST Grade Level 12.6
Powers Sumner Kearl Grade 6.9
Rix Readability 12
Raygor Readability n/a
Fry Readability n/a

Readability Scores:
Readable Rating E
Flesch Reading Ease 23.2
CEFR Level C2
IELTS Level 8+
Spache Score 7.0
New Dale-Chall Score 8.4
Lix Readability 57
Lensear Write 62.9

This instant ai analysis facility made it simpler to adapt the original story into several similar versions, each matching a slightly different audience profile.

(Above) ai generated image using Little People Books 'Dinolad' story as a text prompt

Testing Midjourney and Leonardo.ai for illustrations

Moving beyond text-to-text, we looked at Midjourney and Leonardo.ai for text-to-image. Initially, we found the results were not seamless – on close inspection, every initial image had some flaws. If you look carefully at the leg joints of the dinosaur for example, (image above) you’ll see they are both inaccurate and impossible.

Whilst the ai produced captivating visuals, very few of them matched with the illustration styles of our traditional storytelling book list, but later updates allowed us to upload our own illustrations and get ai to use them as style references.

  (Hand drawn Style Reference)

We concluded that the greatest benefit of generating titles using ai will come from ai’s real time interactive ability to provide each of our readers with their own ‘artisan’ personalised story experience – somewhere between role playing mirror values of their real life situations and experencing ai supplemented augmented reality.

We read a few articles about ai neural link research but nothing suggested to us that these would be able accept externally generated stories. It appears more likely that ai neural links will result in more abstract dream-to-video ai apps. to mirror what the brain does naturally during sleep.

(Above)  "The Earth That Time Forgot" [prompt given to Midjourney ai].

Using ai to create Films:

With our movie director hat on, we tested audio prompts for songs, soundscapes, animations and photorealistic video, leveraging ai to ‘extract’ story from our catalogue of published illustrations and to animate simple scanned-in images into bite sized movie sequences.

As the editing style of most modern film and TV progs. is to use quick cuts between rapidly changing scenes, ai productions like THIS using Sora fit the mood of the present market being in common use on platforms like TikTok. Youtube shorts,Instagram and Reels.  As attention spans are shortening, this follows other signs of the trend such as texting and emoji use.

During our research we reviewed many opinions on the future of ai.  The most interesting view came from Max Tegmark, a renowned physicist and ai researcher leading MIT’s ai Lab. Notably:

We humans have built our identity on being Homo Sapiens, the smartest entities around. As we prepare to be humbled by ever smarter machines, I suggest that we rebrand ourselves as Homo Sentiens.”

AI Literacy
Once upon a medieval time when people used the mark of "X", it was because they were simply illiterate, but as human intelligence is always evolving, those who now "X" are simply those who once tweeted on Twitter. Though the use of "X" for Twitter is probably the shortest example of language abbreviation in recent times, Samuel Morse as early as 1836 designed code that reduced the letter E to a mere dot "." (or dit) and the letter T to a dash "-" (or dah) - a very binary design which when extended to the full alphabet paved the way towards the development of binary language, computing and ai.

Our research findings suggest that the widespread promotion of sentence and word “shortnin'” has resulted in greater human illiteracy, amplified by the rapid uptake of ai apps. [applications].  Initially this may seem to be a backwards step but as the use of ai by so many has given it a data set par excellence, any downside is made up by the speed of its number-crunching which is way beyond human capabilities.

(Above) Generated by nightcafe ai from a two word prompt in June 2023, namely, "My Library".

Our publishing imprint
If painting, carving, handwriting and other crafting skills stay afloat into the future, then bookbinding will survive so long as it keeps its head above water.  For the image above we used a two word prompt, ‘my library‘ and you can see that through the ‘eyes’ of AI in June 2023 the resulting image included a library of books, or, if you are ai, a pattern, in perspective, of thin multicoloured strips divided by receding brown lines.

Like an impressionist painter,  text-to-image ai apps mix ‘hues’ of visual and textual data in such a way that it makes a link in our minds. In the first week of our research we regularly changed text prompts before getting the results we wanted.  This took time but when you consider that ai regularly generates images in seconds that would take weeks to do manually, the few minutes used in re-writing prompts were of no concern.

Will WOW use AI to generate images for Little People Books?
There is one ai app that dropped a few months after our research period concluded that looks perfect for us and that’s a sketching app using predictive ai as you draw. This app also takes text prompts. From the vid. we watched, it interacts realtime with sketched line input so it’s like having a helping hand. I’ll post something about it after we’ve tried it.

Though using ai is nothing like working with an illustrator or author, it is an interesting tool and it can produce stunningly good quality images.

Output, output, output:
Throughout the course of our 4-month study, we achieved remarkable milestones, including the creation of over 50 captivating stories, the preparation of 10 engaging ebooks, the successful publication of 3 ebooks, and the launch of 2 print-on-demand books and colouring books.

Over the same 4-months however, thousands of other groups around the World used ai and produced a similar sized output. Print-on-demand houses were flooded with uploads and most needed to adopt some ai automation to cope with the deluge of raw, quick edited, low production value publications.

(Above) Text Prompt: "Robot surrounded by network of linked icons in space"
The potential of ai in reshaping the landscape of children’s publishing.

Our journey was not without its challenges. While AI has revolutionized the creative process, it has created its own hurdles – ones that still demand human attention. Foremost among these challenges is the need to look carefully at AI generated images and check them for digital idiosyncrasies – the sort that a human illustrator would never make.

Though ai came to its own rescue with ‘inpainting’, our regular photoshop retouch skills remain vital for us.

Ai does not replicate the nuanced understanding and emotional depth that human authors bring to their work, but it does generate a vast variety of content, and faster than any group of humans.

As one example, when we needed the wording for a specific legal document, ai produced the ideal document in seconds which saved our charity both time and funds.

(Above)  Text Prompt: "Busy Street Scene in New Media City"

Use of a Motion Capture Library:
With CCTV cameras and screens filling up our city high streets, their captured content has become of value to ai that generates motion-picture quality movement and settings.  We reviewed high res. ai generated video footage that was minutes long and photo-realistic. Some of the examples were from charities – some footage depicting so-called existing problems, rather than filmed real life situations captured by an actual human film crew.

Using AI in Exhibition Designs:
We got some curious results in this area of ai generation – themed display ideas that our designers may never imagine!  But ‘not in the detail‘, because if you look very closely at the following ai image of exhibits for example, you’ll see that AI has fudged the detail and not understood the picture it is generating from the text prompt we gave it. Take a close look –  it’s un-real, hard to construct, but it’s interesting nevertheless.

Whereas, in our Museum of

Re-Design, Decomposition and Decay

all the exhibits have had a practical life – this marine mooring for example:

and this post under a seafront sign:


Ways that ai might help us when designing WOW exhibitions:
Our exhibitions on marine and freshwater subjects have historically been mounted all over the UK on tour, For our 2026/27 season, we plan to film, edit, then publish them directly on our YouTube channel. Streaming (most apt for a water charity) comes later.

Without a small film crew and basic filming and editing equipment we’re hoping we can use some ai in the form of audio-to-video,  voiceovers and music scores.

Feedback
Despite the various challenges, our study of ai yielded overwhelmingly positive results, supported by customer feedback from focus groups. Parents and educators alike lauded the educational value and engaging storytelling found within our ai-generated storybooks.

Moreover, our efforts to tackle pressing issues such as environmental sustainability and personal well-being were met with resounding acclaim, underscoring the profound impact of literature in shaping young minds and fostering a sense of social responsibility, regardless of the techniques and tools used in the making and the telling.

In conclusion, our first foray into ai-powered publishing has been nothing short of transformative, reaffirming our commitment to innovation and excellence in children’s literature.
As we chart a course for the future, we remain steadfast in our dedication to harnessing the power of ai to inspire, educate, and empower the next generation of readers.


About Us

Little People Books is dedicated to igniting the curiosity and passion for knowledge in children. With a range of captivating materials, we offer an exciting journey that not only educates but also empowers young minds to understand the world around them.

As an imprint of the educational UK charity, World of Water (WoW), Little People Books holds a deep commitment to positive education. We firmly believe that by instilling a sense of wonder and understanding in our children, we can shape attitudes and foster a desire to improve the quality of life now and for the future.

By delving into engaging stories and media, we inspire a profound appreciation for the environment – clean air, clean land, and clean water – essential for a healthier and happier global community.

Invitation:

Join us as we embark on this important mission to create a brighter future.
With our integrated materials, we aim to empower parents, teachers, and carers alike, equipping them with the tools to nurture young minds and guide children towards a deeper understanding of the wonders of life and its potential.

Together, we can inspire a generation of compassionate, informed, and environmentally-conscious individuals who are ready to make a positive impact on the world.

Update ref:  2024.000159/2